The DPT-1 trial, whose clinical component is sponsored by the National Institutes of Health, will initially test the hypothesis that parenteral insulin therapy can prevent type I diabetes amongst cytoplasmic islet cell antibody (ICA) positive first degree relatives. As a basis for this trial, more than 60,000 first degree relatives of patients with type I diabetes will be screened for ICA. ICA positive relatives will be staged and randomized to experimental insulin therapy or close observation (parenteral trial) or oral insulin versus placebo. This proposal will take advantage of both this massive and unique screening program and advances in determination of autoantibodies reacting with biochemically defined recombinant autoantigens. The current proposal will evaluate the ability of such "biochemical" screening to predict type I diabetes in order to understand the natural history of pre-type I diabetes, probe the immunogenetics of autoantibody expression, and develop potential autoantibody surrogate markers of response to therapy. We hypothesize that the presence of a single autoantibody will have high sensitivity, and multiple autoantibodies (eg. equal to or greater than) reduced but excellent sensitivity and a high positive predictive value for diabetes. In addition we hypothesize that in the absence of biochemical autoantibodies ICA will have low positive predictive value. Support for determination of"biochemically" determined autoantibodies, followup and evaluation of ICA negative relatives expressing autoantibodies and matched relatives, and immunogenetic analysis (not part of current DPT- 1 funding) are sought with three centers dividing autoantibody determination: anti- insulin at the Joslin Center, anti-GAD in Gainesville, and anti-ICA512 at the Barbara Davis Center. Support is requested specifically for a natural history study comparing biochemical testing to ICA, IVGTT, HLA DQ alleles, and progression to diabetes. Of note, even the randomized non-treated (parenteral trial) or placebo treated (oral antigen trial of DPT) will exceed in numbers current "natural" history studies.